基于模糊贝叶斯网的危害性分析方法

针对传统的故障模式、影响与危害性分析(FMECA)方法不足的问题,提出了一个基于模糊贝叶斯网的危害性分析方法。该方法将模糊理论与贝叶斯网推理技术结合起来,用三角模糊数来描述专家的模糊评分值；通过模糊集合映射,将其转化为评级的模糊子集；以置信结构的模糊规则,表示故障模式的属性与危害度之间的关系；利用贝叶斯网络推理算法综合置信结构的模糊规则,通过贝叶斯网推理得到模糊子集形式的危害度,再经过去模糊计算,得到故障危害等级的清晰值,从而确定故障模式的危害程度。实验结果表明,所提方法能够提高传统分析方法的准确性和应用范围。

Criticality analysis method based on fuzzy Bayesian networks

Considering the defects of traditional Failure Modes,Effect and Criticality Analysis (FMECA), a criticality analysis method based on fuzzy Bayesian networks was proposed. This approach combined the fuzzy theory with Bayesian network techniques, and fuzzy judgments of experts were described using triangular fuzzy numbers which were transformed into forms of fuzzy subsets of ranking through mapping of fuzzy sets. The fuzzy rules with belief structure were used to represent the relationship between the properties and hazards of the failure modes. The Bayesian network inference algorithms were used to synthesize the fuzzy rules of belief structure, and the hazard degree in the form of fuzzy subsets was obtained by Bayesian inference, through defuzzification calculation, a precise value of fault hazard ranking was gained to determine the hazard degree of the failure mode. The experimental results show that the proposed method is able to improve the accuracy and application range of the traditional analysis method.